Skip to main content
Internet Archive's 25th Anniversary Logo

Full text of "On the Spectroscopic Diversity of Type Ia Supernovae"

See other formats


On the Spectroscopic Diversity of Type la Supernovae 

Kazuhito Hatano^, David Branch^, Eric J. Lentz^, E. Baron^, Alexei V. Filippenko^, and 

Peter M. Garnavich'^ 

Q ■ Received ; accepted 








^Department of Physics and Astronomy, University of Oklahoma, Norman, Oklahoma 

^Department of Astronomy, University of California, Berkeley, CA 94720-3411 
^Department of Physics, University of Notre Dame, Notre Dame, IN 46556 



A comparison of the ratio of the depths of two absorption features in the 
spectra of Type la supernovae (SNe la) near the time of maximum brightness 
with the blueshift of the deep red Si II absorption feature 10 days after maximum 
shows that the spectroscopic diversity of SNe la is multi-dimensional. There is 
a substantial range of blueshifts at a given value of the depth ratio. We also 
find that the spectra of a sample of SNe la obtained a week before maximum 
brightness can be arranged in a "blueshift sequence" that mimics the time evolu- 
tion of the pre-maximum-light spectra of an individual SN la, the well observed 
SN 1994D. Within the context of current SN la explosion models, we suggest 
that some of the SNe la in our sample were delayed-detonations while others 
were plain defiagrations. 

Subject headings: hydrodynamics — radiative transfer - supernovae: general - 
supernovae: individual (SN 1984A, SN 1992A, SN 1994D) 

1. Introduction 

Advancing our understanding of the diversity among Type la supernovae (SNe la) is 
important for identifying the nature of the progenitor binary systems and the explosion 
mechanisms, as well as for boosting confidence in using SNe la as empirical distance 
indicators for cosmology. To first order, SNe la can be divided into normal events that 
are highly but not perfectly homogeneous, and peculiar events like the powerful SN 1991T 
(e.g., Filippenko et al. 1992a) and the weak SN 1991bg (e.g., Filippenko et al. 1992b). 
The next approximation is to regard SNe la as a one-parameter sequence of events, using 
a photometric observable such as Attt-is, the decline of the B magnitude during the first 


15 days after the time of maximum brightness (Philhps 1993; Philhps et al. 1999; Suntzeff 
et al. 1999), or using a spectroscopic observable such as 7^(Si II), the ratio of the depths of 
two absorption features near 5800 and 6100 A that ordinarily are attributed to Si II A5972 
and Si II A6355, respectively (Nugent et al. 1995). The Amis and 72.(Si II) parameters are 
observed to be tightly correlated (Nugent et al. 1995; Garnavich et al. 2000), and because 
they also correlate with peak absolute magnitude they can be regarded as measures of the 
explosion strength. A one-parameter description of SNe la is useful, and in the context of 
Chandrasekhar-mass explosions it can be interpreted in terms of a variation in the mass of 
ejected ^^Ni — although, on the basis of explosion models the actual situation is expected 
to be somewhat more complex than a one-parameter sequence (e.g., Hofiich & Khokhlov 
1996; H5flich et al. 1996). 

Indeed, a one-parameter sequence does not completely account for the observational 
diversity of SNe la. Hamuy et al. (1996) found that some light curves having similar values 
of Ami5 show significant differences of detail, and it has been shown that a two-parameter 
luminosity correction using the B — V color and Attt-is is both necessary and (so far) 
sufficient to standardize SNe la to a common luminosity (Tripp 1998; Tripp & Branch 
1999; Parodi et al. 2000). It also has been noted that some SNe la with normal-looking 
spectra (i.e., having only lines of the usual ions) have exceptionally high blueshifts of their 
absorption features (Branch 1987), and Wells et al. (1994) found a lack of correlation 
between the blueshift of the Si II A6355 absorption near the time of maximum brightness 
and the Amis parameter, in a small sample of well observed SNe la (see also Patat 
et al. 1996). In this Letter we extend the discussion by showing more clearly that the 
spectroscopic diversity among SNe la is multi-dimensional. 



As a measure of the blueshift of a SN la spectrum we adopt the parameter used by 
Branch & van den Bergh (1993; hereafter BvdB), which we denote as 'yio(Si II): it is the 
velocity that corresponds to the blueshift of the flux minimum of the deep absorption feature 
near 6100 A, attributed to Si II A6355, 10 days after the time of maximum brightness. 
We use the data of BvdB, as well as additional data for the following events: SN 1984A 
(Barbon, lijima, & Rosino 1989); SN 1989B (Wells et al. 1994); SN 1991T (Fisher et al. 
1999); SNe 1992A and 1998bu (Jha et al. 1999); SN 1994D (Patat et al. 1996; Filippenko 
1997); SN 1997br (Li et al. 1999); and SN 1999by (Bonanos et al. 2000). As another 
spectroscopic observable we use the 7^(Si II) parameter. Most of the 7^(Si II) values that 
we use are from Garnavich et al. (2000). For events for which they give several slightly 
different values of 7^(Si II), measured at several epochs, we use the median value. For a 
few other events we have estimated 7^(Si II) from published spectra that are referenced 
by BvdB (1993), following the definition of 7?,(Si II) by Nugent et al. (1995). For another 
nine events for which spectra suitable for measuring 7?,(Si II) are unavailable, we have 
used the observed value of Amis (Phillips 1993; Schaefer 1996, 1998; Riess et al. 1999a; 
Suntzeff et al. 1999) to "predict" the value of 7^(Si II), with the help of the tight relation 
between 7?,(Si II) and Arrii^ that is displayed by Garnavich et al. (2000). The characteristic 
uncertainties are 500 km s~^ in i'io(Si II) and 0.04 in 7?,(Si II), which are not negligible but 
not large enough to strongly affect our conclusions. 


3. Results 

3.1. 7^(Si II) versus t;io(Si II) 

The 7^(Si II) parameter is plotted against vio(Si II) in Fig. 1. If SNe la did behave 
as a simple one-parameter sequence, with the mass of ejected ^^Ni being the fundamental 
physical parameter, we would expect 7^(Si II) to decrease smoothly as 'yio(Si II) increases. 
The more powerful the event, the higher the velocity at the outer edge of the core of 
iron-peak elements, and therefore the higher the velocity of the silicon-rich layer. For the 
sample of Fig. 1 taken as a whole, this obviously is not the case. Of course, it is possible that 
the peculiar weak SNe 1991bg, 1999by, and 1986G, and/or the peculiar powerful SNe 1991T 
and 1997br, comprise physically distinct subgroups of SNe la. It is interesting that when we 
disregard these, and restrict our attention to the spectroscopically normal SNe la, it is not 
even clear that there is any correlation between 7^(Si II) and Vio(Si II) at all. Normal SNe la 
that have similar values of 7?.(Si II) can have substantially different values of Vio(Si II), 
so a one-parameter sequence cannot completely account for the spectroscopic diversity of 
normal SNe la. If we were able to use blueshifts measured at a fixed time after explosion, 
rather than at a fixed time with respect to maximum brightness. Fig. 1 would look slightly 
different because of the diversity in SN la rise times from explosion to maximum brightness 
(Riess et al. 1999b), but this would not change our conclusion about the need for more 
than a one-dimensional spectroscopic sequence. 

3.2. Pre— maximum spectra 

Now we turn to the diversity among pre-maximum-light spectra of SNe la. Fig. 2 
shows spectra of five events, all obtained about one week before the time of maximum 
brightness, arranged in a "blueshift sequence." Some things to notice in Fig. 2 are that (1) 


as is to be expected, the absorption features are broader in the higher-blueshift events; (2) 
the absorption features tend to be deeper in the higher-blueshift events (not necessarily 
expected); and (3) the Ca II absorption appears as a single feature in the high-blueshift 
SNe 1984A and 1992A, while it appears split in the lower blueshift SNe 1989B, 1994D, and 

For comparison with Fig. 2, Fig. 3 shows a time series of pre-maximum-light spectra 
of the particularly well observed SN 1994D. The degree to which Fig. 3 resembles Fig. 2 
is intriguing. The —12 day spectrum of SN 1994D in Fig. 3 looks much like the —7 day 
spectrum of SN 1984A in Fig. 2; the SN 1994D absorptions are at their deepest at the 
earliest times when the blueshifts are the highest; and the Ca II absorption of SN 1994D 
evolves from unsplit to split. It is almost as if normal SNe la follow a standard pattern 
of pre-maximum spectroscopic evolution after all, but their spectra are not in phase 
with respect to the time of maximum brightness. (This statement should not be taken 
too literally, and certainly it cannot be said of the spectroscopically peculiar SNe la; 
there was no pre-maximum phase at which the spectrum of SN 1994D resembled that of 
SN 1991bg-like or SN 1991T-like events.) 

Fig. 4 shows some comparisons of observed pre-maximum SN la spectra with synthetic 
spectra that have been generated with the parameterized spectrum-synthesis code SYNOW. 
[Brief descriptions of the code can be found in recent applications of SYNOW, e.g., Fisher 
et al. (1999), Millard et al. (1999), Hatano et al. (1999a); details are in Fisher (2000).] The 
top panel compares the —6 day spectrum of SN 1994D with a best-fit synthetic spectrum 
that has a velocity at the photosphere of Vphot= 11,000 km s^^ and includes lines of O I, 
Mg II, Si II, Si III, S II, Ca II, and Fe II. The second panel compares the —6 day spectrum 
of SN 1992A with two synthetic spectra. In one of them (dashed line) the SYNOW input 
parameters are the same as in the top panel, except that Vphd has been increased from 


11,000 to 13,000 km s~^; this produces about the right absorption-feature blueshifts for 
SN 1992A, but most of the synthetic absorptions are too weak. In the other synthetic 
spectrum of the middle panel (dotted line), the optical depths of the O I, Si II, and S II 
lines have been increased to achieve a better fit. Similarly, in the bottom panel the —7 
day spectrum of SN 1984A is compared to two synthetic spectra. In one of them (dashed 
line) only Vphot has been changed, from 13,000 to 17,000 km s^^ to get the right blueshifts. 
In the other (dotted line), the optical depths of the Mg II, Si II, Ca II, and Fe II lines 
have been further increased [and a weak high-velocity Fe II component (cf. Hatano et al. 
1999a) has been introduced] to make the synthetic absorptions strong enough. Fig. 4 shows 
that these rather different pre-maximum spectra of SNe 1994D, 1992A, and 1984A can be 
matched well with lines of the same ions, but with different photospheric velocities and line 
optical depths. In general, the higher the blueshift of the spectrum, the higher the line 
optical depths. 

4. Discussion 

Attempting an in-depth interpretation of the causes of the spectroscopic diversity 
illustrated in this Letter would be premature, but a brief discussion is in order. As stressed 
above, the main point is that spectroscopically normal SNe la that have similar values of 
7^(Si II) (or Amis) have a significant range in fio(Si II), which is not what we would expect 
if SNe la behaved as a simple one-parameter sequence of events. One possible explanation 
is that two or more explosion mechanisms are involved. In this regard it is noteworthy 
that Lentz et al. (2000) find from detailed NLTE calculations that some of the published 
delayed-detonation models can account for the very high Si II blueshift of SN 1984 A. It 
may be that the seven events ranging diagonally from SN 1992A to SN 1983G in Fig. 1 
are delayed detonations having a range of strengths, while the other spectroscopically 

normal events are plain deflagrations. From this point of view it seems useful to regard the 
7?,(Si II) axis as a measure of the ejected mass of ^^Ni and the Vio(Si II) axis as a measure 
of the amount of matter ejected at high velocity; the latter is much higher in delayed 
detonations than in plain deflagrations (Lentz et al. 2000). It is not clear whether the weak 
SN 1991bg-like events should be regarded as weak delayed detonations, weak deflagrations, 
or something else. For suggestions that the powerful SN 1991T-like events may come from 
super-Chandrasekhar white dwarf merger products, see Fisher et al. (1999). 

We should keep in mind that the distribution of events in Fig. 1 also could be 
affected by asymmetries. BvdB (1993) argued, on the basis of a perceived connection 
between 'L'io(Si II) and parent-galaxy type, that the whole range of Vio(Si II) values cannot 
be attributed entirely to asymmetries, and SNe la in general are observed to have low 
polarization (Wang et al. 1996), but the possibility that asymmetry does play some role in 
Fig. 1 is not excluded. The ejected matter could be inherently asymmetric because of, e.g., 
a deflagration-detonation transition that begins at a single point rather than on a spherical 
shell (Livne 1999), and/or an asymmetry could be imposed by the presence of a donor star 
(Marietta et al. 2000). 

Figs. 2, 3, and 4 raise some additional questions. Why are the line optical depths 
higher when the blueshifts are higher? In the Sobolev approximation, line optical depths 
are proportional to the time since the explosion (owing to the decreasing velocity gradient 
and the increasing size of the line resonance region), and this time dependence is only 
partially offset by the decreasing density at the photosphere. Therefore, temperature 
evolution seems to be required to account for the decreasing line strengths that we see in 
Fig. 3. If the temperature of the line-forming layers of SN 1994D increased during the rise 
towards maximum brightness, [as indicated by broad-band photometry (Richmond et al. 
1995; Patat et al. 1996; Meikle et al. 1996)], the Ca II and Si II optical depths would have 


decreased as the optical depths of the weaker Si III and Fe III hnes increased (Hatano et al. 
1999b); this does seem to have been the case for SN 1994D (cf. Hatano et al. 1999a). 

Temperature differences may also be partly responsible for the blueshift sequence of 
Fig. 2. The resemblance of the spectrum of SN 1994D at —12 days to that of SN 1984A 
at —7 days may mean that the high-velocity layers of SN 1984A at —7 days were still as 
cool as those of SN 1994D at —12 days. The higher densities in the high-velocity layers of 
delayed detonations are likely to cause them to be cooler than those of plain deflagrations 
at a fixed time after explosion. 

Our suggestion that Fig. 1 is populated by both delayed detonations and deflagrations 
seems to be a reasonable working hypothesis within the context of current SN la explosion 
models, but there are many uncertainties associated with such models (Hillebrandt & 
Niemeyer 2000). Our understanding of the causes of the spectroscopic diversity of SNe la 
is still in a rudimentary state. Many more high-quality observed spectra and much more 
spectroscopic analysis are needed. 

We thank Darrin Casebeer, Dean Richardson, and Thomas Vaughan for helpful 
discussions. This work was supported by NSF grants AST-9417213, 9731450, and 9987438, 
as well as by NASA grant NAG5-3505. 



Barbon, R., lijima, T., & Rosino, L. 1989, A&A, 220, 83 

Bonanos, A. et al. 2000, in preparation 

Branch, D. 1987, ApJ, 316, L81 

Branch, D., & van den Bergh, S., 1993, AJ, 105, 2231 (BvdB) 

Fihppenko, A. V. 1997, ARAA, 35, 309 

Fihppenko, A. V. et al. 1992a, ApJ, 384, L15 

Fihppenko, A. V. et al 1992b, AJ, 104, 1543 

Fisher, A. 2000, PhD thesis. University of Oklahoma 

Fisher, A., Branch, D., Hatano, K., & Baron, E., 1999, MNRAS, 304, 67 

Garnavich, P. M. et al. 2000, in preparation 

Hamuy, M. et al. 1996, A J, 112, 2438 

Hatano, K., Branch, D., Fisher, A., Baron, E., & Fihppenko, A. V. 1999a, ApJ, 525, 881 

Hatano, K., Branch, D., Fisher, A., MiUard, J., & Baron, E. 1999b, ApJS, 121, 233 

Hillebrandt, W., & Niemeyer, J. C. 2000, ARAA, in press 

Hdflich, P., & Khokhlov, A. 1996, ApJ, 457, 500 

H5flich, P. et al. 1996, ApJ, 472, L81 

Jha, S. et al., 1999, ApJS, 125, 73 

Kirshner, R. P. et al. 1993, ApJ, 415, 589 

Leibundgut, B. et al., 1993, AJ, 105, 301 

Lentz, E., Baron, E., Branch, D., & Hauschildt, P. 2000, ApJ, in press 

Li, W. D. et al., 1999, AJ, 117, 2709 

- 11- 

Livne, E. 1999, ApJ, 527, 97 

Marietta, E., Burrows, A., & Fryxell, B. 2000, ApJS, in press 

Meikle, W. P. S. et al. 1996, MNRAS, 281, 263 

Millard, J. et al. 1999, ApJ, 527, 746 

Nugent, P. E., Phillips, M. M., Baron, E., Branch, D., & Hauschildt, P. 1995, ApJ, 455, 

Parodi, B. R., Saha, A., Sandage, A., & Tammann, G. A., 2000, ApJ, in press 

Patat, P., Benetti, S., Cappellaro, E., Danziger, I. J., Delia Valle, M., Mazzali, P. M., & 
Turatto, M. 1996, MNRAS, 278, 111 

Phillips, M. M., 1993, ApJ, 413, L105 

Phillips, M. M., Lira, P., Suntzeff, N. B., Schommer, R. A., Hamuy, M., & Maza, J. 1999, 
AJ, 118, 1766 

Richmond, M. W. et al. 1995, AJ, 109, 2121 

Riess, A. G. et al. 1999a, AJ, 117, 707 

Riess, A. G. et al. 1999b, A J, 118, 2675 

Schaefer, B. E., 1996, ApJ, 460, L19 

Schaefer, B. E., 1998, ApJ, 509, 80 

Suntzeff, N. B. et al., 1999, AJ, 117, 1175 

Tripp, R. 1998, A&A, 331, 815 

Tripp, R., & Branch, D. 1999, ApJ, 525, 209 

Wang, L., Wheeler, J. C., Li, Z., & Clocchiatti, A. 1996, ApJ, 467, 435 

Wegner, G., & McMahan, R. K. 1987, AJ, 93, 287 


Wells, L. A. et al., 1994, AJ, 108, 2233 

This manuscript was prepared with the AAS IAT[t;X macros v5.0. 


Fig. 1: The 7^(Si II) parameter is plotted against fio(Si II). Arrows denote spectroscopically 
peculiar SNe la. Open symbols mean that 7^(Si II) has been obtained from a relation 
between Arrii^ and 7^(Si II). 

Fig. 2: Spectra of five SNe la about a week before the time of maximum light (SN 1984A: 
Wegner & McMahan 1987; SN 1992A: Kirshner et al. 1993; SN 1989B: Wells et al. 1994; 
SN 1994D: Filippenko 1997; SN 1990N: Leibundgut et al. 1993) are arranged in a blueshift 
sequence. The vertical lines, drawn as an aide to the eye, are blueshifted by 15,000 km s~^ 
from A6355 (Si II) and A3945 (Ca II). The vertical displacements are arbitrary. 

Fig. 3: A time series of pre-maximum spectra of SN 1994D (Patat et al. 1996; Filippenko 
1997, and unpublished). The vertical lines are the same as in Fig. 2. 

Fig. 4: Comparisons of observed spectra (solid curves) of SN 1994D at —6 days (top panel), 
SN 1992A at —6 days (middle panel), and SN 1984A at —7 days (bottom panel) with 
SYNOW synthetic spectra. See the text for descriptions of the synthetic spectra. 








v5 0.4 




► 83R 










74G ^90N 97br 
95D o 37C 

91T n 95E 









v,„(SiII) (lO' kms"') 

Fig 2 


S -15 







4000 5000 6000 7000 8000 

Rest Wavelength (Angstroms) 


Fig 3 

3 - 

a ^ 





3000 4000 

5000 6000 7000 8000 9000 10000 
Rest Wavelength (Angstroms) 

Fig 4 


4000 5000 6000 7000 8000 

Rest Wavelength (Angstroms)